Track planning method and device and vehicle

Abstract

The embodiment of the invention provides a trajectory planning method and device and a vehicle. According to the method, the state information of the feature points in the planning time domain of the vehicle can be determined according to the predicted trajectory of the traffic participants around the vehicle; and according to the state information of the feature points in the own vehicle planning time domain, solving the trajectory in the own vehicle planning time domain. According to the invention, the algorithm link of the trajectory planning algorithm is simplified, the executable trajectory can be quickly decided, and the calculation cost, resource consumption and volume production carrying cost of the trajectory planning algorithm are reduced.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology, and more specifically, to a trajectory planning method, device, and vehicle. Background Technology

[0002] In vehicle interaction scenarios, current trajectory planning methods first employ complex processing techniques to analyze the game between the vehicle and other traffic participants and generate a decision result (which is not an executable trajectory). This decision result then undergoes further processing to plan an executable trajectory. The long algorithmic chain of current trajectory planning methods results in high computational costs, significant resource consumption, and high costs for mass production and deployment. Summary of the Invention

[0003] This application proposes a trajectory planning method, device, and vehicle, which can directly determine the state information of the vehicle's feature points based on the interaction between the vehicle and other traffic participants, and solve the trajectory based on the state information of the vehicle's feature points. This simplifies the algorithm chain of the trajectory planning algorithm, can quickly determine the executable trajectory, and reduces the computational cost, resource consumption, and mass production cost of the trajectory planning algorithm, thus solving the above-mentioned technical problems.

[0004] In a first aspect, embodiments of this application provide a trajectory planning method, which includes: determining the state information of feature points in the planning time domain of the vehicle based on the predicted trajectories of traffic participants around the vehicle; and solving the trajectory in the planning time domain of the vehicle based on the state information of the feature points in the planning time domain.

[0005] Secondly, embodiments of this application provide a trajectory planning device, which includes: a feature point determination module, used to determine the state information of feature points in the planning time domain of the vehicle based on the predicted trajectories of traffic participants around the vehicle; and a trajectory solving module, used to solve the trajectory in the planning time domain of the vehicle based on the state information of the feature points in the planning time domain of the vehicle.

[0006] Thirdly, embodiments of this application provide an electronic device, which includes a memory and a processor. The memory stores an application program that, when invoked by the processor, causes the processor to execute the method provided in the embodiments of this application.

[0007] Fourthly, this application provides a vehicle that includes the electronic equipment provided in the third aspect.

[0008] Fifthly, embodiments of this application provide a computer-readable storage medium storing program code, which, when invoked by a processor, causes the processor to execute the method provided in embodiments of this application.

[0009] Sixthly, embodiments of this application provide a computer program product, which, when invoked by a processor, causes the processor to execute the method provided in embodiments of this application.

[0010] The trajectory planning method provided in this application has the following technical effects: it can directly determine the state information of the vehicle's feature points based on the interaction between the vehicle and other traffic participants, and solve the trajectory based on the state information of the vehicle's feature points. This simplifies the algorithm chain of the trajectory planning algorithm, can quickly determine the executable trajectory, and reduces the computational cost, resource consumption, and mass production cost of the trajectory planning algorithm. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments and drawings obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0012] Figure 1 This is a schematic flowchart of a trajectory planning method provided in an embodiment of this application; Figure 2 This is a flowchart illustrating a trajectory planning method provided in another embodiment of this application; Figure 3 This is a schematic flowchart illustrating the process of solving the trajectory corresponding to a lane-keeping scenario according to an embodiment of this application; Figure 4 This is a flowchart illustrating the process of solving the trajectory corresponding to a lane-changing scenario according to an embodiment of this application; Figure 5 This is a flowchart illustrating the process of solving the trajectory corresponding to an obstacle avoidance scenario within the road, according to an embodiment of this application. Figure 6 This is a structural block diagram of a trajectory planning device provided in an embodiment of this application; Figure 7 This is a structural block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0013] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0014] The trajectory planning method in this application can be applied to a trajectory planning device or a vehicle, and the trajectory planning device can be deployed in a vehicle. The vehicle is equipped with a sensing module for sensing environmental information around the vehicle. The sensing device can include, but is not limited to, at least one of a camera, lidar, millimeter-wave radar, and ultrasonic radar. The vehicle can include, but is not limited to, gasoline vehicles or new energy vehicles, and new energy vehicles can include electric vehicles, which can include, but are not limited to, pure electric vehicles, hybrid electric vehicles, or fuel cell vehicles.

[0015] See Figure 1 , Figure 1 This is a schematic flowchart of a trajectory planning method provided in an embodiment of this application. The trajectory planning method may include steps S110-S120.

[0016] Step S110: Determine the state information of feature points in the planning time domain of the vehicle based on the predicted trajectories of traffic participants around the vehicle.

[0017] Traffic participants surrounding the vehicle can include other vehicles in the vicinity. Within each planning cycle, the predicted trajectories of traffic participants can be obtained through the vehicle's prediction module. For details on obtaining the predicted trajectories of traffic participants, please refer to relevant technologies; this application will not elaborate further. The trajectory points on the predicted trajectories of traffic participants contain their state information, which may include, but is not limited to: the absolute position information (latitude and longitude), heading angle, speed, acceleration, yaw rate, type, and size parameters of the traffic participants.

[0018] The environmental information surrounding the vehicle can include structured road environment information within a certain range around the vehicle. This environmental information may include, but is not limited to, the number of lanes, the guidance attributes of each lane, navigation information, lane width, weather conditions, and the current absolute position information (latitude and longitude) of traffic participants, heading angle, speed, acceleration, yaw rate, type, and size parameters. Within each planning cycle, the vehicle's perception module can acquire the environmental information surrounding the vehicle in real time.

[0019] The planning time domain refers to the time span of the planned trajectory; it can also be understood as the planning period for the vehicle's trajectory planning. In this application, the planning time domain can range from 8 to 10 seconds. The planning time domain can include multiple moments (each moment corresponding to a decision segment). For example, the planning time domain can be divided into moments by second. Assuming the planning time domain is 10 seconds, dividing it by second would result in the planning time domain comprising 10 moments.

[0020] A feature point is a point on the planned trajectory, and each moment in the planning time domain corresponds to a feature point. The state information of a feature point may include, but is not limited to: the relative position of the vehicle with respect to the predicted trajectory of other traffic participants, heading angle, speed, acceleration, etc.

[0021] Within each planning cycle, the vehicle's current status information can be obtained through the vehicle's chassis data. This current status information may include, but is not limited to: the current moment, the vehicle's absolute position information (latitude and longitude), heading angle, speed, acceleration, yaw rate, historical trajectory point list, and navigation path road environment data.

[0022] Based on the upstream sensing process, the sensing information for the current planning cycle (i.e., the sensing information for the current frame) can be updated. This sensing information can include environmental information surrounding the vehicle and predicted trajectories of traffic participants. The acquired sensing information can then be preprocessed. Preprocessing may include, but is not limited to, validity checks and coordinate system transformation. Coordinate system transformation refers to converting the environmental information surrounding the vehicle and the predicted trajectories of traffic participants to the SLT coordinate system. The SLT coordinate system adds a time axis T to the Frenet coordinate system (two-dimensional). The SLT coordinate system is three-dimensional. In the SLT coordinate system, S represents the vehicle's distance traveled on the road (also called its longitudinal position). The L-axis represents the vehicle's lane position or lateral offset on the road. T represents time, indicating the vehicle's position at a specific point in time during planning.

[0023] Based on the predicted trajectories of traffic participants, "characteristic moments" of interaction between the vehicle and traffic participants can be identified to determine several vehicle state information to form "feature points". For example, for each moment in the vehicle's planning time domain (i.e., each moment sampled at a time resolution within the planning time domain, such as 1 second), the state information of the vehicle's feature points at that moment can be determined based on the state information of the trajectory points on the predicted trajectories of traffic participants.

[0024] As one implementation method, the time required for the vehicle to reach the preset speed limit can be determined based on the vehicle's current speed and the preset speed limit; the maximum distance the vehicle can reach within the planning time domain can be determined based on the vehicle's current speed, the preset speed limit, the preset acceleration, and the time required for the vehicle to reach the preset speed limit; the difference between the longitudinal position of the traffic participant in front of the vehicle and the longitudinal following distance can be determined; the endpoint (longitudinal position) of the planning time domain can be determined based on the maximum distance or the difference; and the state information of the feature points at each moment within the planning time domain can be determined based on the predicted trajectory of the traffic participant between the vehicle's current position and the endpoint.

[0025] The time taken for the vehicle to reach the preset speed limit refers to the time t_Vmax taken for the vehicle to go from its current speed Vc to the preset speed limit Vmax, where t_Vmax = (Vmax - Vc) / a, and a is a pre-defined fixed value, for example, a = 0.8 m / s². The farthest distance the vehicle can reach within the planning time domain is S_max = Vc·t_Vmax + 0.5a·t_Vmax^2 + Vmax·(T - t_Vmax).

[0026] In this application, scenarios can be classified according to vehicle behavior and environmental information. Scenarios may include, but are not limited to, lane keeping scenarios, lane changing scenarios, and obstacle avoidance scenarios. For different scenarios, the planning time domain endpoint can be determined according to the predicted trajectory of traffic participants and in a specific way corresponding to the scenario. Based on the current position of the vehicle, the planning time domain endpoint, and the predicted trajectory of traffic participants, the state information of feature points at each moment in the planning time domain can be determined.

[0027] As one implementation method, for lane keeping scenarios, if the farthest distance that the vehicle can reach within the planning time domain is less than or equal to the difference between the longitudinal position of the traffic participant in front of the vehicle and the longitudinal following distance, then the endpoint (longitudinal position) of the planning time domain is determined based on the farthest distance; if the farthest distance is greater than the difference, then the endpoint (longitudinal position) of the planning time domain is determined based on the difference.

[0028] As one implementation method, for lane-keeping scenarios, feature points and corresponding state information for each moment in the vehicle's planning time domain are set based on the state information (including relative position, heading angle, speed, acceleration, etc.) of feature points on the predicted trajectory of traffic participants at each moment. For example, the state information of feature points can be set according to the SLT coordinate system. For instance, assuming the planning time domain includes 10 moments (T=1, 2, ..., 10), the L coordinate of the feature point can be determined based on the state information of the trajectory points of the traffic participants at the corresponding moment, the T coordinate is 1, 2, ..., 10, and the S coordinate is between 0 and the longitudinal position of the vehicle's planning time domain endpoint. For lane-keeping scenarios, the longitudinal position of the vehicle's planning time domain endpoint can be determined based on the starting point S_obs_front of the traffic participants' predicted trajectory (i.e., the current longitudinal position of the vehicle in front) and the farthest distance S_max that the vehicle can reach within the planning time domain. For example, the difference between the current longitudinal position S_obs_front of the vehicle in front and the longitudinal following distance D_save can be calculated. The smaller of this difference and the farthest distance that the vehicle can reach within the planning time domain can be determined as the longitudinal position S_LK of the vehicle at the end of the planning time domain, where S_LK = min(S_max, S_obs_front - D_save).

[0029] As one implementation method, for lane keeping scenarios, based on the predicted trajectory of traffic participants from the current position of the vehicle to the destination, the feature points and corresponding status information of the vehicle at each moment can be set according to the status information of the trajectory points at each moment on the predicted trajectory of the traffic participants (including relative position, heading angle, speed, acceleration, etc.), thereby obtaining the status information of the feature points at each moment in the planning time domain of the vehicle (i.e., the feature point sequence).

[0030] As one implementation method, for lane-changing scenarios or obstacle-avoidance scenarios, the endpoint (longitudinal position) of the planning time domain is determined based on the farthest distance that the vehicle can reach within the planning time domain; if the trajectory of the lane-changing scenario or obstacle-avoidance scenario cannot be solved based on the endpoint determined based on the farthest distance, the endpoint (longitudinal position) of the planning time domain is re-determined based on the difference between the longitudinal position of the traffic participant in front of the vehicle and the longitudinal following distance.

[0031] In some embodiments, for lane-changing scenarios, before determining the difference between the longitudinal position of the traffic participant in front of the vehicle and the longitudinal following distance, traffic participants (i.e., target traffic participants) that interact with the vehicle in the planning time domain can be selected based on the minimum allowable lane-changing gap_t, the minimum lane-changing distance, and the farthest distance S_max that the vehicle can reach in the planning time domain at each time point.

[0032] As one implementation method, based on the predicted trajectory of traffic participants from the current position of the vehicle to the planning time domain endpoint, the feature points and corresponding status information of the vehicle at each moment can be set according to the status information (including relative position, heading angle, speed, acceleration, etc.) of the trajectory points at each moment on the predicted trajectory of each traffic participant (or each target traffic participant), thereby obtaining the status information (i.e., feature point sequence) of the feature points at each moment in the planning time domain of the vehicle for each traffic participant (or each target traffic participant).

[0033] As one implementation method, for obstacle avoidance scenarios within the lane, lateral distance dilation (the dilation distance can be the lateral safety distance between vehicles) can be performed on the trajectory points of each traffic participant's predicted trajectory at each moment to obtain the dilated trajectory of each traffic participant; the lateral and longitudinal position information of each traffic participant's dilated trajectory encroaching on the lane where the vehicle is located at each moment can be determined; based on the lateral and longitudinal position information of each traffic participant's dilated trajectory encroaching on the lane where the vehicle is located at each moment, the state information of the feature points at the corresponding moment in the planning time domain can be determined.

[0034] In some embodiments, after determining the state information of the feature points at each time step in the planning time domain, the feature points can be filtered according to the maximum expected acceleration a_exp_max (for example, filtering out feature points in the feature point sequence whose acceleration exceeds the maximum expected acceleration) to obtain effective feature points, thereby reducing the amount of data and improving the calculation speed of subsequent trajectory solving.

[0035] Step S120: Solve the trajectory of the vehicle in the planning time domain based on the state information of the feature points in the planning time domain.

[0036] In one implementation, step S120 may include: calculating the trajectories corresponding to multiple scenarios based on the time taken for the vehicle to reach the preset speed limit, the farthest distance the vehicle can reach within the planning time domain, and the state information of feature points within the planning time domain, thus obtaining multiple trajectories; and selecting the optimal trajectory from the multiple trajectories as the trajectory within the vehicle's planning time domain. The multiple scenarios may include, but are not limited to, lane keeping scenarios, lane changing scenarios, and obstacle avoidance scenarios, and the trajectories corresponding to each of these scenarios can be calculated separately to obtain multiple trajectories.

[0037] As one implementation method, for lane keeping scenarios, if the farthest distance that the vehicle can reach within the planning time domain is less than or equal to the difference between the longitudinal position of the traffic participant in front of the vehicle and the longitudinal following distance, then the trajectory is solved using a first trajectory solving method based on the farthest distance that the vehicle can reach within the planning time domain, the time taken for the vehicle to reach the preset speed limit, and the state information of the feature points within the planning time domain of the vehicle; if the farthest distance that the vehicle can reach within the planning time domain is greater than the difference, then the trajectory is solved using a second trajectory solving method based on the difference and the state information of the feature points within the planning time domain of the vehicle.

[0038] As one implementation method, for lane-changing scenarios or obstacle-avoidance scenarios, the trajectory can be solved using the first trajectory solving method based on the farthest distance that the vehicle can reach within the planning time domain, the time it takes for the vehicle to reach the preset speed limit, and the state information of the feature points within the planning time domain of the vehicle.

[0039] As another implementation method, for lane-changing or obstacle-avoidance scenarios, the trajectory can first be solved using a first trajectory solving method based on the farthest distance the vehicle can reach within the planning time domain, the time it takes for the vehicle to reach the preset speed limit, and the state information of feature points within the vehicle's planning time domain. Then, collision detection and trajectory quality assessment are performed on the solved trajectory. If the solved trajectory passes the collision detection and trajectory quality assessment, it can be directly determined as the trajectory for the lane-changing or obstacle-avoidance scenario. If the solved trajectory fails the collision detection and trajectory quality assessment, a second trajectory solving method can be used to solve the trajectory based on the difference between the longitudinal position of the traffic participants in front of the vehicle and the longitudinal following distance, as well as the state information of feature points within the vehicle's planning time domain.

[0040] As one implementation method, the trajectory can be solved by using the first trajectory solving method based on the farthest distance that the vehicle can reach in the planning time domain, the time it takes for the vehicle to reach the preset speed limit, and the state information of the feature points in the planning time domain. This can include steps S1A-S2A.

[0041] Step S1A: If the time t_Vmax taken for the vehicle to reach the preset speed limit is greater than or equal to the planned time domain T, then the trajectory is solved based on the current state information of the vehicle and the state information of the endpoint of the planned time domain.

[0042] As one implementation method, if t_Vmax≥T, a trajectory planning algorithm can be used to solve the trajectory based on the vehicle's current state information state_start and the planned time-domain endpoint state information state_goal_T. The trajectory planning algorithm uses a seventh-order state equation, which eliminates the need to construct an optimization problem, resulting in high execution efficiency. The generated trajectory satisfies second-order kinematic continuity, facilitating trajectory quality evaluation in subsequent steps.

[0043] Step S2A: If the time taken for the vehicle to reach the preset speed limit is less than the planned time domain, then based on the time taken for the vehicle to reach the preset speed limit, obtain the state information of the trajectory connection points from the state information of the feature points within the vehicle's planned time domain; solve the first half of the trajectory based on the vehicle's current state information and the state information of the trajectory connection points; solve the second half of the trajectory based on the state information of the trajectory connection points and the state information of the endpoint; merge the first half of the trajectory and the second half of the trajectory.

[0044] As one implementation method, if t_Vmax < T, the characteristic time t_a0 of the trajectory connection point can be obtained by first rounding down to the nearest second based on the time t_Vmax taken by the vehicle to reach the preset speed limit. Then, the state information of the characteristic point at that time can be obtained from the state information of the characteristic point within the planning time domain of the vehicle based on t_a0, thus obtaining the state information of the trajectory connection point. A trajectory planning algorithm can be used to solve the first half of the trajectory (i.e., the trajectory from the current time of the vehicle to time t_a0) based on the current state information of the vehicle and the state information of the trajectory connection point. The second half of the trajectory (i.e., the trajectory from time t_a0 to time T) can be solved based on the state information of the trajectory connection point and the state information of the end point of the planning time domain. Finally, the first half of the trajectory and the second half of the trajectory are merged into the entire trajectory. The trajectory planning algorithm is a seventh-order state equation. Solving the trajectory using a seventh-order state equation does not require constructing an optimization problem, has high execution efficiency, and the generated trajectory satisfies second-order kinematic continuity, which is easy for subsequent steps to evaluate the trajectory quality.

[0045] As one implementation method, the trajectory can be solved by using the second trajectory solution method based on the difference between the longitudinal position of the traffic participants in front of the vehicle and the longitudinal following distance, as well as the state information of the feature points in the vehicle's planning time domain. This can include steps S1B-S4B.

[0046] Step S1B: Based on the current state information of the vehicle and the state information of the feature points within the planning time domain of the vehicle, solve for multiple first-half trajectories.

[0047] As one implementation method, a trajectory planning algorithm can be used to solve for multiple first-half trajectories based on the current state information of the vehicle and the state information of feature points within the vehicle's planning time domain. The trajectory planning algorithm uses a seventh-order state equation, which eliminates the need to construct an optimization problem, resulting in high execution efficiency. The generated trajectories satisfy second-order kinematic continuity, facilitating trajectory quality evaluation in subsequent steps.

[0048] Step S2B: Through collision detection and trajectory quality evaluation, the optimal first half trajectory is selected from the multiple first half trajectories, and the state information of the end feature points of the optimal first half trajectory is obtained.

[0049] Collision detection filters out trajectories with collision risks in the first half of the trajectory. Trajectory quality assessment selects the optimal trajectory from those without collision risks. The terminal feature point of the optimal trajectory is the connection point between the first and second halves of the trajectory (i.e., the feature point at time t_a0). For collision detection algorithms, please refer to relevant technologies. In some embodiments, trajectory quality assessment can use a pre-trained quality assessment model (e.g., a deep learning model) to score trajectories without collision risks, and the trajectory with the highest score is determined as the optimal trajectory. Factors considered in the scoring may include, but are not limited to, the acceleration and velocity of feature points, and the slope of interpolation of state information between pairs of feature points (those with large slopes are discarded). For specific model structure selection and model training for the quality assessment model, please refer to relevant technologies.

[0050] Step S3B: Solve the second half of the trajectory based on the state information of the end feature point and the state information of the endpoint.

[0051] A trajectory planning algorithm can be used to solve the second half of the trajectory based on the state information of the terminal feature points and the state information of the endpoint. The trajectory planning algorithm uses a seventh-order state equation, which eliminates the need to construct an optimization problem, resulting in high execution efficiency. The generated trajectory satisfies second-order kinematic continuity, facilitating trajectory quality evaluation in subsequent steps.

[0052] Step S4B: Merge the first half of the trajectory and the second half of the trajectory.

[0053] After obtaining the first half and the second half of the trajectory, the first half and the second half of the trajectory can be merged into the whole trajectory.

[0054] After obtaining the trajectories corresponding to multiple scenarios, the optimal trajectory can be selected from these trajectories to serve as the trajectory for the vehicle's planning time domain. Specifically, the effectiveness, collision detection, and trajectory quality assessments can be performed on multiple trajectories respectively, and the trajectory with the highest quality can be selected from the valid trajectories without collision risk as the trajectory for the vehicle's planning time domain. The trajectory planning method of this application can quickly solve for the most efficient and safest (optimal for a single frame) trajectory within the current planning cycle. Considering the interaction between the vehicle and various traffic participants, the algorithm chain of the trajectory planning algorithm can be simplified, and an executable trajectory can be quickly determined, reducing the computational cost, resource consumption, and mass production deployment cost of the trajectory planning algorithm.

[0055] For each trajectory in the various scenarios, the validity of the trajectory can be determined based on the relative relationship between the predicted trajectory of the traffic participant and the trajectory itself. If the trajectory is valid, it is retained; otherwise, it is filtered out. Parameters used to determine validity may include, but are not limited to: the traffic participant's orientation angle, lateral position offset, the positional relationship between the vertices of the dimension box and the lane boundary, the degree of intrusion of the dimension box into the target lane, and the lateral speed.

[0056] Specifically, for each trajectory in the trajectories corresponding to multiple scenarios, collision detection can be performed on the trajectory. If the collision detection result indicates that there is a collision risk, the trajectory is filtered out; if the collision detection result indicates that there is no collision risk, the trajectory is retained.

[0057] For each trajectory within the various scenarios, a trajectory quality assessment can be performed, yielding an assessment result (e.g., a quality score). The parameters used for trajectory quality assessment may include, but are not limited to, trajectory efficiency parameters and parameters related to the matching degree between trajectory behavior and navigation direction. For example, trajectory efficiency equals the weighted average of the velocities of all feature points on the trajectory. The matching degree between trajectory behavior and navigation direction equals the trajectory behavior direction value × the navigation information direction value × the weight. It is understandable that if multiple parameters are calculated, the minimum value of all parameters can be taken as the trajectory quality assessment result.

[0058] After effectiveness assessment and collision detection, the trajectory with the best evaluation result (e.g., the highest quality score) can be selected from the remaining valid trajectories that do not pose a collision risk and used as the trajectory for the vehicle's planning time domain (i.e., the trajectory for the current planning cycle).

[0059] After obtaining the optimal trajectory within the planning time domain of the autonomous vehicle, the trajectory can be transformed from the SLT coordinate system to the autonomous vehicle coordinate system (the autonomous vehicle coordinate system refers to a coordinate system with the center point of the rear axle of the vehicle as the origin, the forward direction of the vehicle as the X-axis, the left side of the vehicle as the Y-axis, and the vertical upward direction as the Z-axis) to form an executable trajectory for the autonomous vehicle. Then, the trajectory is sent to the downstream control link for control execution.

[0060] Steps S110-S120 have the following technical effects: the state information of the vehicle's feature points is directly determined based on the interaction between the vehicle and other traffic participants, and the trajectory is solved based on the state information of the vehicle's feature points. This simplifies the algorithm link of the trajectory planning algorithm, enables the rapid determination of an executable trajectory, and reduces the computational cost, resource consumption, and mass production cost of the trajectory planning algorithm.

[0061] See Figure 2 , Figure 2This is a flowchart illustrating a trajectory planning method provided in another embodiment of this application. The trajectory planning method may include steps S210-S250.

[0062] Step S210: Update the perception information of the current frame and preprocess the perception information.

[0063] Step S220: Determine the time it takes for the vehicle to reach the preset speed limit and the farthest distance the vehicle can reach in the current frame.

[0064] It should be understood that the farthest distance that the vehicle can reach in the current frame is the farthest distance that the vehicle can reach within the planning time domain mentioned above.

[0065] Step S230: Based on the state information of trajectory points at each moment on the predicted trajectory of traffic participants interacting with the vehicle, the feature points and their state information at the corresponding moments in the vehicle planning time domain for different scenarios are determined using a method corresponding to different scenarios.

[0066] Step S240: For each of the multiple scenarios, use the feature points and state information of the feature points corresponding to the scenario to solve the trajectory corresponding to the scenario.

[0067] Step S250: Perform collision detection and trajectory quality evaluation on the trajectories of multiple scenes respectively, and select the optimal trajectory from the trajectories of multiple scenes as the executable trajectory of the current frame.

[0068] Steps S210-S250 have the following technical effects: the state information of the vehicle's feature points is directly determined based on the interaction between the vehicle and other traffic participants, and the trajectory is solved based on the state information of the vehicle's feature points. This simplifies the trajectory planning algorithm chain, enables the rapid determination of executable trajectories, reduces the computational cost and resource consumption of the algorithm, and allows the method to be deployed on low-computation platforms, enabling wider application and lower-cost mass production solutions.

[0069] See Figure 3 , Figure 3 This is a schematic flowchart illustrating the process of solving the trajectory corresponding to a lane-keeping scenario according to an embodiment of this application. Solving the trajectory corresponding to a lane-keeping scenario may include steps S310-S3110.

[0070] Step S310: Determine the longitudinal position of the planning time domain endpoint.

[0071] Determine the difference between the current longitudinal position of the vehicle ahead and the longitudinal following distance; determine the farthest distance the vehicle can reach in the current frame (i.e., the farthest distance the vehicle can reach within the planning time domain), and determine the smaller of the difference and the farthest distance as the longitudinal position of the planning time domain endpoint. If the longitudinal position of the planning time domain endpoint is the farthest distance the vehicle can reach in the current frame, then execute steps S320-S340 and S390-S3110. If the longitudinal position of the planning time domain endpoint is the difference between the current longitudinal position of the vehicle ahead and the longitudinal following distance, then execute steps S350-S3110.

[0072] Step S320: If the longitudinal position of the end point of the planned time domain is the farthest distance that the vehicle can reach in the current frame, determine whether the time taken for the vehicle to reach the preset speed limit is less than the planned time domain.

[0073] If the time taken for the vehicle to reach the preset speed limit is less than the planned time domain, the first half of the trajectory and the second half of the trajectory are solved in segments according to the time taken for the vehicle to reach the preset speed limit, and the first half of the trajectory and the second half of the trajectory are merged into the whole trajectory (step S330), and collision detection is performed on the whole trajectory (step S390).

[0074] If the time taken for the vehicle to reach the preset speed limit is not less than the planned time domain, then the trajectory is solved based on the current state information of the vehicle and the state information of the end point of the planned time domain (step S340), and collision detection is performed on the solved trajectory (step S390).

[0075] Step S330: Solve the first half of the trajectory and the second half of the trajectory in segments according to the time taken for the vehicle to reach the preset speed limit, and merge the first half of the trajectory and the second half of the trajectory into the whole trajectory.

[0076] Step S340: Solve the trajectory based on the current state information of the vehicle and the state information of the planned time domain endpoint.

[0077] Step S350: When the longitudinal position of the end point of the planning time domain is the difference between the current longitudinal position of the vehicle in front and the longitudinal following distance, solve several first half trajectories based on the current state information of the vehicle and the state information of the feature points sampled at each time point in the planning time domain according to the time resolution.

[0078] Step S360: Perform collision detection on each first half of the trajectory.

[0079] If any first half of the trajectory passes the collision detection, then the trajectory quality of each first half of the trajectory that passes the collision detection is evaluated, and the first half of the trajectory with the best trajectory quality is obtained (step S370). If no first half of the trajectory passes the collision detection, then the result that the trajectory of the lane keeping scenario cannot be solved is output (step S3110).

[0080] Step S370: Evaluate the trajectory quality of each first half of the trajectory that passes the collision detection, and obtain the first half of the trajectory with the best trajectory quality.

[0081] Step S380: Based on the state information of the end feature points of the optimal first half trajectory and the state information of the planned time domain endpoint, solve the second half trajectory, and merge the optimal first half trajectory and the second half trajectory into the whole trajectory.

[0082] Step S390: Perform collision detection on the entire trajectory.

[0083] If the entire trajectory passes the collision detection, the trajectory for the lane keeping scenario is output (step S3100). If the entire trajectory fails the collision detection, the result that the trajectory for the lane keeping scenario cannot be solved is output (step S3110).

[0084] Step S3100: Output the trajectory of the lane keeping scenario.

[0085] Step S3110: Output the result of the inability to solve the trajectory of the lane keeping scenario.

[0086] A detailed description of steps S310-S3110 can be found in the relevant sections of the foregoing embodiments, and will not be repeated here. Steps S310-S3110 have the following technical effects: the state information of the vehicle's feature points is directly determined based on the interaction between the vehicle and other traffic participants, and the trajectory is solved based on the state information of the vehicle's feature points. This simplifies the trajectory planning algorithm chain in lane-keeping scenarios, enables rapid decision-making of executable trajectories in lane-keeping scenarios, and reduces the computational cost and resource consumption of the algorithm.

[0087] See Figure 4 , Figure 4 This is a flowchart illustrating the process of solving the trajectory corresponding to a lane-changing scenario according to an embodiment of this application. Solving the trajectory corresponding to a lane-changing scenario may include steps S410-S4140.

[0088] Step S410: Determine the furthest distance that the vehicle can reach in the current frame as the longitudinal position of the planning time domain endpoint, and determine the time taken for the vehicle to reach the preset speed limit.

[0089] Step S420: Determine whether the time taken for the vehicle to reach the preset speed limit is less than the planned time range.

[0090] If the time taken for the vehicle to reach the preset speed limit is less than the planned time domain, the first half of the trajectory and the second half of the trajectory are solved in segments according to the time taken for the vehicle to reach the preset speed limit, and the first half of the trajectory and the second half of the trajectory are merged into the whole trajectory (step S430), and collision detection is performed on the whole trajectory (step S450).

[0091] If the time taken for the vehicle to reach the preset speed limit is not less than the planned time domain, then the trajectory is solved based on the current state information of the vehicle and the state information of the end point of the planned time domain (step S440), and collision detection is performed on the solved trajectory (step S450). Step S430: Solve the first half of the trajectory and the second half of the trajectory in segments according to the time taken for the vehicle to reach the preset speed limit, and merge the first half of the trajectory and the second half of the trajectory into the whole trajectory.

[0092] Step S440: Solve the trajectory based on the current state information of the vehicle and the state information of the planned time domain endpoint.

[0093] Step S450: Perform collision detection on the entire trajectory.

[0094] If the entire trajectory passes the collision detection, the trajectory of the lane-changing scenario is output (step S4130). If the entire trajectory fails the collision detection, the difference between the longitudinal position of the preceding vehicle and the longitudinal following distance is determined as the longitudinal position of the planning time domain endpoint, and traffic participants who need to interact with the vehicle are selected (step S460).

[0095] Step S460: Determine the difference between the longitudinal position of the preceding vehicle and the longitudinal following distance as the longitudinal position of the planning time domain endpoint, and filter out the target traffic participants who need to interact with the vehicle.

[0096] Step S470: Select the state information of feature points at each time point in the vehicle planning time domain based on the predicted trajectory of the target traffic participants.

[0097] Step S480: Solve for several first half trajectories based on the current state information of the vehicle and the state information of the feature points at each time point in the planning time domain.

[0098] Step S490: Perform collision detection on each first half of the trajectory.

[0099] If any first half of the trajectory passes the collision detection, then the trajectory quality of each first half of the trajectory that passes the collision detection is evaluated, and the first half of the trajectory with the best trajectory quality is obtained (step S4100). If no first half of the trajectory passes the collision detection, then the result that the trajectory of the lane keeping scenario cannot be solved is output (step S4140).

[0100] Step S4100: Evaluate the trajectory quality of each first half of the trajectory that passes the collision detection, and obtain the first half of the trajectory with the best trajectory quality.

[0101] Step S4110: Based on the state information of the end feature points of the optimal first half trajectory and the state information of the planned time domain endpoint, solve the second half trajectory, and merge the optimal first half trajectory and the second half trajectory into the whole trajectory.

[0102] Step S4120: Perform collision detection on the entire trajectory.

[0103] If the entire trajectory passes the collision detection, the trajectory of the lane-changing scenario is output (step S4130). If the entire trajectory fails the collision detection, the result that the trajectory of the lane-changing scenario cannot be solved is output (step S4140).

[0104] Step S4130: Output the trajectory of the lane-changing scenario.

[0105] Step S4140: Output the result of the inability to solve the trajectory of the lane-changing scenario.

[0106] A detailed description of steps S410-S4140 can be found in the relevant sections of the foregoing embodiments, and will not be repeated here. Steps S410-S4140 have the following technical effects: the state information of the vehicle's feature points is directly determined based on the interaction between the vehicle and other traffic participants, and the trajectory is solved based on the state information of the vehicle's feature points. This simplifies the trajectory planning algorithm chain for lane-changing scenarios, enables rapid decision-making of executable trajectories for lane-changing scenarios, and reduces the computational cost and resource consumption of the algorithm.

[0107] See Figure 5 , Figure 5 This is a schematic flowchart illustrating the process of solving the trajectory corresponding to an obstacle avoidance scenario within a lane, according to an embodiment of this application. Solving the trajectory corresponding to an obstacle avoidance scenario within a lane may include steps S510-S5140.

[0108] Step S510: Determine the furthest distance that the vehicle can reach in the current frame as the longitudinal position of the planning time domain endpoint, and determine the time taken for the vehicle to reach the preset speed limit.

[0109] Step S520: Determine whether the time taken for the vehicle to reach the preset speed limit is less than the planned time range.

[0110] If the time taken for the vehicle to reach the preset speed limit is less than the planned time domain, the first half of the trajectory and the second half of the trajectory are solved in segments according to the time taken for the vehicle to reach the preset speed limit, and the first half of the trajectory and the second half of the trajectory are merged into the whole trajectory (step S530), and collision detection is performed on the whole trajectory (step S550).

[0111] If the time taken for the vehicle to reach the preset speed limit is not less than the planned time domain, then the trajectory is solved based on the current state information of the vehicle and the state information of the end point of the planned time domain (step S540), and collision detection is performed on the solved trajectory (step S550). Step S530: Solve the first half of the trajectory and the second half of the trajectory in segments according to the time taken for the vehicle to reach the preset speed limit, and merge the first half of the trajectory and the second half of the trajectory into the whole trajectory.

[0112] Step S540: Solve the trajectory based on the current state information of the vehicle and the state information of the planned time domain endpoint.

[0113] Step S550: Perform collision detection on the entire trajectory.

[0114] If the entire trajectory passes the collision detection, the trajectory of the obstacle avoidance scenario in the lane is output (step S5130). If the entire trajectory fails the collision detection, the difference between the longitudinal position of the preceding vehicle and the longitudinal following distance is determined as the longitudinal position of the planning time domain endpoint, and traffic participants that need to interact with the vehicle are selected (step S560).

[0115] Step S560: Determine the difference between the longitudinal position of the preceding vehicle and the longitudinal following distance as the longitudinal position of the planning time domain endpoint, and filter out the traffic participants (i.e., target traffic participants) that need to interact with the vehicle.

[0116] Step S570: Select the state information of feature points at each time point in the vehicle planning time domain based on the predicted trajectory of the target traffic participants.

[0117] Step S580: Solve for several first half trajectories based on the current state information of the vehicle and the state information of the feature points at each time point in the planning time domain.

[0118] Step S590: Perform collision detection on each first half of the trajectory.

[0119] If any first half of the trajectory passes the collision detection, then the trajectory quality of each first half of the trajectory that passes the collision detection is evaluated, and the first half of the trajectory with the best trajectory quality is obtained (step S5100). If no first half of the trajectory passes the collision detection, then the result that the trajectory of the obstacle avoidance scenario in the path cannot be solved is output (step S5140).

[0120] Step S5100: Evaluate the trajectory quality of each first half of the trajectory that passes the collision detection, and obtain the first half of the trajectory with the best trajectory quality.

[0121] Step S5110: Based on the state information of the end feature points of the optimal first half trajectory and the state information of the planned time domain endpoint, solve the second half trajectory, and merge the optimal first half trajectory and the second half trajectory into the whole trajectory.

[0122] Step S5120: Perform collision detection on the entire trajectory.

[0123] If the entire trajectory passes the collision detection, the trajectory of the obstacle avoidance scenario within the lane is output (step S5130). If the entire trajectory fails the collision detection, the result that the trajectory of the obstacle avoidance scenario within the lane cannot be solved is output (step S5140).

[0124] Step S5130: Output the trajectory of the obstacle avoidance scene within the lane.

[0125] Step S5140: Output the result of the inability to solve the trajectory of the obstacle avoidance scenario within the track.

[0126] A detailed description of steps S510-S5140 can be found in the relevant parts of the foregoing embodiments, and will not be repeated here. Steps S510-S5140 have the following technical effects: the state information of the vehicle's feature points is directly determined based on the interaction between the vehicle and other traffic participants, and the trajectory is solved based on the state information of the vehicle's feature points. This simplifies the trajectory planning algorithm link for obstacle avoidance scenarios within the lane, enables rapid decision-making of executable trajectories for obstacle avoidance scenarios within the lane, and reduces the computational cost and resource consumption of the algorithm.

[0127] See Figure 6 , Figure 6 This is a structural block diagram of a trajectory planning device provided in an embodiment of this application. The trajectory planning device 100 may include a feature point determination module 110 and a trajectory solving module 120. The feature point determination module 110 is used to determine the state information of feature points in the vehicle's planning time domain based on the predicted trajectories of traffic participants around the vehicle. The trajectory solving module 120 is used to solve the trajectory in the vehicle's planning time domain based on the state information of the feature points in the vehicle's planning time domain. For specific details of the operations performed, please refer to the embodiments of the trajectory planning method of this application.

[0128] Those skilled in the art will clearly understand that the apparatus provided in the embodiments of this application can implement the methods provided in the embodiments of this application. The specific working process of the described apparatus and modules can be found in the corresponding processes of the methods in the embodiments of this application, and will not be repeated here.

[0129] In the embodiments provided in this application, the coupling, direct coupling, or communication connection between the modules shown or discussed may be indirect coupling or communication coupling through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms. The embodiments of this application do not impose specific limitations on this.

[0130] Furthermore, the functional modules in the embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0131] See Figure 7 , Figure 7This is a structural block diagram of an electronic device provided in an embodiment of this application. The electronic device 200 may include a memory 210 and a processor 220. The memory 210 stores an application program, which is configured to cause the processor 220 to execute the method provided in the embodiment of this application when invoked by the processor 220.

[0132] Processor 220 may include one or more processing cores. Processor 220 uses various interfaces and lines to connect to various parts of the entire electronic device 200, and is used to run or execute instructions, programs, code sets or instruction sets stored in memory 210, as well as to call and run or execute data stored in memory 210, and perform various functions of electronic device 200 and process data.

[0133] The processor 220 can be implemented using at least one of the following hardware forms: Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor 220 can integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the displayed content; and the modem handles wireless communication. It is understood that the modem can also be implemented separately as a communication chip, without being integrated into the processor 220.

[0134] The memory 210 may include random access memory (RAM) or read-only memory (ROM). The memory 210 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 210 may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for implementing at least one function, instructions for implementing the various method embodiments described above, etc. The data storage area may store data created by the electronic device 200 during use.

[0135] This application also provides a vehicle that includes electronic equipment 200.

[0136] This application also provides a computer-readable storage medium. The computer-readable storage medium stores program code that, when invoked by a processor, causes the processor to execute the method provided in this application.

[0137] Computer-readable storage media can be electronic storage devices such as flash memory, electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), hard disk, or ROM.

[0138] In some embodiments, the computer-readable storage medium includes a non-volatile computer-readable storage medium (Non-TCRSM). The computer-readable storage medium has storage space for program code that performs any of the method steps described above. This program code can be read from or written to one or more computer program products. The program code may be compressed in an appropriate form.

[0139] This application also provides a computer program product, which includes a computer program that, when invoked by a processor, causes the processor to execute the method provided in the embodiments of this application.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A trajectory planning method characterized by, include: Based on the predicted trajectories of traffic participants around the vehicle, determine the state information of feature points within the vehicle's planning time domain; Based on the state information of feature points within the vehicle's planning time domain, the trajectory within the vehicle's planning time domain is solved.

2. The method of claim 1, wherein, The process of determining the state information of feature points within the vehicle's planning time domain based on the predicted trajectories of traffic participants around the vehicle includes: Based on the vehicle's current speed and the preset speed limit, determine the time it will take for the vehicle to reach the preset speed limit; Based on the vehicle's current speed, preset speed limit, preset acceleration, and the time it takes for the vehicle to reach the preset speed limit, determine the farthest distance the vehicle can reach within the planned time domain. Determine the difference between the longitudinal position of the traffic participant in front of the vehicle and the longitudinal following distance; The endpoint of the planning time domain is determined based on the farthest distance or the difference. Based on the predicted trajectory of the traffic participants from the current position of the vehicle to the destination, the state information of the feature points at each moment in the vehicle planning time domain is determined.

3. The method of claim 2, wherein, Determining the endpoint of the planning time domain based on the farthest distance or the difference includes: For lane keeping scenarios, if the farthest distance is less than or equal to the difference, the endpoint of the planning time domain is determined based on the farthest distance. If the farthest distance is greater than the difference, then the endpoint of the planning time domain is determined based on the difference.

4. The method of claim 2, wherein, Determining the endpoint of the planning time domain based on the farthest distance or the difference includes: For lane-changing scenarios or obstacle-avoidance scenarios within the lane, the endpoint of the planning time domain is determined based on the farthest distance; If the trajectory of a lane-changing scenario or an obstacle-avoiding scenario cannot be solved based on the endpoint determined by the farthest distance, then the endpoint of the planning time domain is re-determined based on the difference.

5. The method of claim 2, wherein, The step of solving the trajectory within the self-planning time domain based on the state information of feature points within the self-planning time domain includes: Based on the time taken for the vehicle to reach the preset speed limit, the farthest distance the vehicle can reach within the planning time domain, and the state information of the feature points within the planning time domain, the trajectories corresponding to multiple scenarios are solved, resulting in multiple trajectories. The optimal trajectory is selected from the multiple trajectories and used as the trajectory for the vehicle's planning time domain.

6. The method of claim 5, wherein, The process involves calculating the trajectories for multiple scenarios based on the time taken for the vehicle to reach the preset speed limit, the maximum distance the vehicle can reach within the planning time domain, and the state information of feature points within the planning time domain. This includes: For lane keeping scenarios, if the farthest distance that the vehicle can reach within the planning time domain is less than or equal to the difference, then the trajectory is solved using the first trajectory solving method based on the farthest distance that the vehicle can reach within the planning time domain, the time taken for the vehicle to reach the preset speed limit, and the state information of the feature points within the planning time domain of the vehicle. If the farthest distance that the vehicle can reach within the planning time domain is greater than the difference, then the trajectory is solved using the second trajectory solution method based on the difference and the state information of the feature points within the planning time domain of the vehicle.

7. The method of claim 5, wherein, The process involves calculating the trajectories for multiple scenarios based on the time taken for the vehicle to reach the preset speed limit, the maximum distance the vehicle can reach within the planning time domain, and the state information of feature points within the planning time domain. This includes: For a lane-changing scene or an in-lane obstacle-avoiding scene, a first trajectory solving manner is used to solve a trajectory according to a farthest distance that a host vehicle can reach within a planning time domain, a time consumed by the host vehicle to reach a preset speed limit, and state information of feature points within the planning time domain of the host vehicle.

8. The method of claim 7, wherein, After the first trajectory solving manner is used to solve a trajectory according to a farthest distance that a host vehicle can reach within a planning time domain, a time consumed by the host vehicle to reach a preset speed limit, and state information of feature points within the planning time domain of the host vehicle for a lane-changing scene or an in-lane obstacle-avoiding scene, the method further comprises: collision detection and trajectory quality evaluation are performed on the solved trajectory; if the solved trajectory fails to pass the collision detection and the trajectory quality evaluation, a second trajectory solving manner is used to solve a trajectory according to the difference and the state information of the feature points within the planning time domain of the host vehicle.

9. The method according to claim 6 or 7, characterized in that, The first trajectory solving manner comprises: if the time consumed by the host vehicle to reach the preset speed limit is greater than or equal to the planning time domain, a trajectory is solved according to current state information of the host vehicle and state information of an end point of the planning time domain; if the time consumed by the host vehicle to reach the preset speed limit is less than the planning time domain, state information of a trajectory connection point is obtained from the state information of the feature points within the planning time domain of the host vehicle according to the time consumed by the host vehicle to reach the preset speed limit, a first-half trajectory is solved according to current state information of the host vehicle and the state information of the trajectory connection point, a second-half trajectory is solved according to the state information of the trajectory connection point and the state information of the end point, and the first-half trajectory and the second-half trajectory are merged.

10. The method of claim 6 or 8, wherein, The second trajectory solving manner comprises: a plurality of first-half trajectories are solved according to current state information of the host vehicle and the state information of the feature points within the planning time domain of the host vehicle; an optimal first-half trajectory is selected from the plurality of first-half trajectories through collision detection and trajectory quality evaluation, and state information of an end feature point of the optimal first-half trajectory is obtained; a second-half trajectory is solved according to the state information of the end feature point and the state information of the end point; the first-half trajectory and the second-half trajectory are merged.

11. An electronic device, comprising: comprises: a memory and a processor, the memory storing an application program, the application program being configured to cause the processor to execute the method according to any one of claims 1-10 when the application program is invoked by the processor.

12. A vehicle characterized by comprising: comprises: the electronic device according to claim 11.

Images